BriefGPT.xyz
Jul, 2018
组合生成对抗网络:学习图像条件下的二元组合
Compositional GAN: Learning Conditional Image Composition
HTML
PDF
Samaneh Azadi, Deepak Pathak, Sayna Ebrahimi, Trevor Darrell
TL;DR
本文提出 Composition-by-Decomposition 网络,采用两个独立分布的物体生成具有真实纹理和形状的合成图像,以此捕捉多物体之间的空间相互作用关系,从而在生成场景方面获得更好的效果。
Abstract
generative adversarial networks
(GANs) can produce images of surprising complexity and realism, but are generally modeled to sample from a single latent source ignoring the explicit
spatial interaction
between mu
→